Optimizing Frames for Search Engines
Dale Goetsch
Search Innovation
November 13, 2002
revised May 12, 2003
Background
Because of the way framed web pages are created, search engine robots have a difficult time spidering sites built in frames. As a general rule, search engine robots are not very good at executing client-side code, and framed pages are "built" on the client side. The best way to make a website accessible to the robots is to take it out of frames, but what can be done if the site absolutely must remain in frames?
How frames are built
Typically the "framing" page--the page that includes the <FRAMESET> tags--does not contain any links to the rest of the website; rather, it contains only information necessary for the browser to construct the framed pages.
<html>
<head>
<title>SuperWidget XYZ from XYZ</title>
</head>
<frameset rows="20%,80%">
<frame name="frameTop" src="header.html">
<frameset cols="30%,70%">
<frame name="frameLeft" src="navbar.html">
<frame name="frameMain" src="page1.html">
</frameset>
</frameset>
</html>
Figure 1—index.html: the "framing" page
The framing page loads the files named in the <FRAME> tags into the frames defined in the framing page. In our example, they look like this:
<html> <head> <title>Navbar</title> </head> <body> <a href="page1.html" target="frameMain">Page 1</a><br> <a href="page2.html" target="frameMain">Page 2</a><br> <a href="page3.html" target="frameMain">Page 3</a><br> </body> </html>
Figure 2—navbar.html
<html> <head> <title>Page 1</title> </head> <body> Welcome to XYZ, home of the new and improved SuperWidget XYZ. We have the best widgets available anywhere today, and at half the price of most leading widgets! </body> </html>
Figure 3—page1.html
What if you can't do frames?
Unfortunately, most robots cannot navigate through this page. They do not understand the <FRAME> tags, and are unable to move through this page to the pages "navbar.html" or "page1.html". Without being able to move through here, there is literally nothing of interest for the robot to index, so there will really be no information in a search engine listing, if the site is listed at all.
The <NOFRAMES> section
One of the tricks that was incorporated into HTML with the advent of frames was the recognition that a page may be accessed by older browsers that are incapable of rendering framed pages: they literally cannot understand the <FRAMESET> tag. This is why there also exists the <NOFRAMES> tag. This then allows users on browsers that are not frames-enabled to at least see something on a website. Typically, the <NOFRAMES> section is wasted on a message telling the user to get a newer browser, thus:
<html>
<head>
<title>SuperWidget XYZ from XYZ</title>
</head>
<frameset rows="20%,80%">
<frame name="frameTop" src="header.html">
<frameset cols="30%,70%">
<frame name="frameLeft" src="navbar.html">
<frame name="frameMain" src="page1.html">
</frameset>
</frameset>
<noframes>
This web site must be viewed using a frames-capable web browser. Your
web browser, however, is not capable of displaying frames.
</noframes>
</html>
Figure 4—index.html with added <NOFRAMES> section
This page now has information that the robot can spider and include in the search engine database. Unfortunately, you will now be known as the website with the content:
This web site must be viewed using a frames-capable web browser. Your web browser, however, is not capable of displaying frames.
Figure 5—search engine listing with poor <NOFRAMES> text
Text in <NOFRAMES> section
This is probably not what you want your potential visitors to see when they look in the search engine listings. Since the robot can spider at least this page, it only makes sense to put your best foot forward and put some real content into the <NOFRAMES> section. That way, your search engine listing will actually tell something about your site, rather than just annoy people because they choose to use a browser that doesn't do frames.
<html>
<head>
<title>SuperWidget XYZ from XYZ</title>
</head>
<frameset rows="20%,80%">
<frame name="frameTop" src="header.html">
<frameset cols="30%,70%">
<frame name="frameLeft" src="navbar.html">
<frame name="frameMain" src="page1.html">
</frameset>
</frameset>
<noframes>
Welcome to XYZ, home of the new and improved SuperWidget XYZ. We
have the best widgets available anywhere today, and at half the price
of most leading widgets!
</noframes>
</html>
Figure 6—index.html with better <NOFRAMES> text
This is a dramatic improvement, because we now have real content on the page that the robot can read and include in the search engine database. You have now upgraded your search engine listing to this:
Welcome to XYZ, home of the new and improved SuperWidget XYZ. We have the best widgets available anywhere today, and at half the price of most leading widgets!
Figure 7—search engine listing with better <NOFRAMES> text
Unfortunately, this is still not enough.
Navigation in <NOFRAMES> section
Most websites comprise multiple pages. Even though you may have a lot of navigation links in your navbar.html file, the robot will never see it. That means that if you want the robot to crawl the rest of your site, you will need to give it some links in the <NOFRAMES> section of the page. To duplicate the navbar functionality, you will need to add those links to the <NOFRAMES> text, like this:
<html>
<head>
<title>SuperWidget XYZ from XYZ</title>
</head>
<frameset rows="20%,80%">
<frame name="frameTop" src="header.html">
<frameset cols="30%,70%">
<frame name="frameLeft" src="navbar.html">
<frame name="frameMain" src="page1.html">
</frameset>
</frameset>
<noframes>
Welcome to XYZ, home of the new and improved SuperWidget XYZ. We
have the best widgets available anywhere today, and at half the price
of most leading widgets!<br>
<a href="page1.html" target="frameMain">Page 1</a><br>
<a href="page2.html" target="frameMain">Page 2</a><br>
<a href="page3.html" target="frameMain">Page 3</a><br>
</noframes>
</html>
Figure 8—index.html with added hyperlinks
Now you have the best of both worlds: you have text that the robot can grab, and you also have links that the robot can follow to access the rest of your site. As long as you have links to all of the pages on your website that you want the robot to access, you are home free now, search engine-wise. The robot follows the link to the file "page2.html", for example, and indexes the text on that page. How useful this newly-indexed content is to your visitor is now up to you.
Why is site framed?
People use framed sites for a number of reasons: ease in navigation, uniform appearance throughout a site, keeping your company name front-and-center, and so on. In other words, there is probably a reason why you wanted to display the pages on your site (file1.html, file2.html, file3.html) within the frames designated in the "framing" page. Your search engine entries, however, will not keep the pages in this configuration--remember that the robot didn't do frames, so the search engine database knows nothing of frames now either. That means the hyperlink created in the search engine listing will load only the individual page (file2.html), and not put it in its overall context. That's not what you wanted, or you would have designed the site that way!
Loading page into frames
In order to force the user's browser to load a given page into the framed environment that you wanted, you must employ some JavaScript sleight-of-hand. Specifically, you need to make each page aware that it wants to load only within the frames that you have designed. This is a two-step process that involves placing some JavaScript code in each page on the site.
Individual pages
For each of the individual pages, you need to add an awareness whether they are loaded into a frame, or sitting by themselves as an individual document in the browser window. This is accomplished by adding the following JavaScript to the page, typically within the HEAD section:
<script>
<!--
function frameMe(){
//name of the frameset page (be sure to add path info)
var frameset = "index.html";
//name of the frame to load document into
var frame = "frameMain";
page = new String(self.document.location.pathname);
if(window.name != frame){
var contents = frameset + "?" + page + "&" + frame;
top.location.replace(contents);
}
return;
}
frameMe();
//-->
</script>
Figure 9—JavaScript for individual pages
You will replace some of the parameters here with names more appropriate to your situation:
Let's implement this for the page "page1.html", which we want loaded into the frame named "right" that is defined in the framing page "index.html". Remember to add your meta tags and give the page a meaningful title. Note the substitutions referenced above.
<html>
<head>
<title>Page 1</title>
<meta content="description" value="SuperWidget and all other widgets at half price from XYZ">
<meta content="keywords" value="SuperWidget XYZ best widgets half price widgets">
<script>
<!--
function frameMe(){
//name of the frameset page (be sure to add path info)
var frameset = "index.html";
//name of the frame to load document into
var frame = "frameMain";
page = new String(self.document.location.pathname);
if(window.name != frame){
var contents = frameset + "?" + page + "&" + frame;
top.location.replace(contents);
}
return;
}
frameMe();
//-->
</script>
</head>
<body>
<h1>Page 1</h1>
Welcome to XYZ, home of the new and improved SuperWidget XYZ. We
have the best widgets available anywhere today, and at half the price
of most leading widgets!
</body>
</html>
Figure 10—individual page with JavaScript added
The "framing" page
For the "framing" page, there are two additional pieces of JavaScript that must be added. The first is a JavaScript function that is placed in the <HEAD> section of the page:
<script>
<!--
function loadDoc(){
//an array containing the querystring portion of the URL
// of the page being loaded into this frame
var query = window.location.search.split('&');
//the first member of that array (without the "?"): the page name
var page = query[0].substr(1);
//the second member of that array: the frame name
var frame = query[1];
if(page && frame){
//only do the replacement if there was a querystring to parse
var contents = "top." + frame + ".location.replace('" + page + "')";
eval(contents);
}
return;
}
//-->
</script>
Figure 11—JavaScript for "framing" page
There are no substitutions in this code. Place it exactly as shown.
The other piece of code that must be placed is an "onLoad" event handler, that is placed in the <FRAMESET> tag, like this:
<frameset cols="30%,70%" onLoad="loadDoc()">
Figure 12—onLoad code for "framing" page
Remember that your frameset may have "cols" or "rows" attributes, and the numbers may vary. When we have it all together, the completed "framing" page will now look like this:
<html>
<head>
<title>SuperWidget XYZ from XYZ</title>
<script>
<!--
function loadDoc(){
//an array containing the querystring portion of the URL
// of the page being loaded into this frame
var query = window.location.search.split('&');
//the first member of that array (without the "?"): the page name
var page = query[0].substr(1);
//the second member of that array: the frame name
var frame = query[1];
if(page && frame){
//only do the replacement if there was a querystring to parse
var contents = "top." + frame + ".location.replace('" + page + "')";
eval(contents);
}
return;
}
//-->
</script>
</head>
<frameset rows="20%,80%">
<frame name="frameTop" src="header.html">
<frameset cols="30%,70%" onLoad="loadDoc()">
<frame name="frameLeft" src="navbar.html">
<frame name="frameMain" src="page1.html">
</frameset>
</frameset>
<noframes>
Welcome to XYZ, home of the new and improved SuperWidget XYZ. We
have the best widgets available anywhere today, and at half the price
of most leading widgets!<br>
<a href="page1.html" target="frameMain">Page 1</a><br>
<a href="page2.html" target="frameMain">Page 2</a><br>
<a href="page3.html" target="frameMain">Page 3</a><br>
</noframes>
</html>
Figure 13—all components placed in "framing" page
Completion
Once you have made these changes for the "framing" page and all of the individual pages, you are now ready for the search engine robots to visit your site. They will be able to access all of the pages in your site, and when your visitors click on your listing in the search engines, your individual pages will load in the way you designed them to work.
|
Dale Goetsch is the Technical Consultant for Search Innovation Marketing, a Search Engine Optimization company serving small businesses. He has over fourteen years experience in software development. Along with programming in Perl, JavaScript, ASP and VB, he is adept at technical writing and editing. |
